The purpose of this study was to compare visual and computerized assisted semen analysis (CASA) in an in vitro fertilization program. Both methods were performed on the day of 316 oocyte retrievals. CASA was performed using a Hamilton 7.2 device adjusted in accordance with the settings defined by a group of 14 French centers. Despite a few discrepancies, the correlation between visual and CASA sperm counts was significant (p < 0.001), so that between the two total motility assessments. The same correlation (Pearson's test) was observed between the fertilization rate and the parameters obtained by visual analysis and CASA. VAP, VSL, VCL and ALH exhibited threshold values under which the mean fertilization rate was below 50%. This study demonstrates the satisfactory reliability of CASA although it cannot yet replace visual analysis.